News and Blog Report: Oilers are terrible when Taylor Hall is on the ice

[Ol' Jase]

http://oilersnation.com/2014/1/16/there-is-something-wrong-with-taylor-hall

So many jumps in logic and confirmation bias I just had to post it.

As always, the usual questions arise, most importantly why do people continue to use unit derived metrics to evaluate individual performance? But to outright claim "the Oilers are terrible" when Hall is on the ice in one breath and in the next show the team scoring at EV at an elite level when he is on the ice, and then claim definite proof of Hall "having something wrong"...it's really something else to witness.

Advanced stats had, and still have, so much promise in terms of analysis tools. It's a damn shame that the spotlight is being shone on them by people who simply can not interpret them properly.

Click to read more...

 

[Tad Mikowsky]

More reason why I don't really care for advance statistics.

 

[Mr Forever]

Hockey is an incredibly difficult sport to measure with an advanced stat, because it isn't played in a vacuum like baseball is. I think there is a future for statistics like these, but the context in which they are presented must be sharpened to avoid logic fallacies like this one.

 

[CanadianSuperPromise]

I think you interpreted the article wrong.

 

[StoveTopStauffer]

There is something with Hall? No. There is something with Eakins.

Hall, Eakins 43.2%, Krueger 50.4%
Nugent-Hopkins, Eakins 45.1%, Krueger 49.9%.


Truth be told Corsi is generally better under Eakins but he sure has some outliers in other stat departments.

What else is crazy is how much scoring chances have decreased for all the increase in Corsi.

Scoring chance plus/minus for 15 min ESTOI

Hall, Eakins +2.1, Krueger +3.7
Eberle, Eakins +2.2, Krueger +4.0;
Nugent-Hopkins, Eakins +1.4, Krueger +2.5

Various other players where Krueger is higher. Yak is .1 higher for Eakins. Either way it isn't a small reduction and they certainly aren't better.

 

[backhandsauce]

Jonathon Willis is an idiot.

 

[Beerfish]

Advanced stats are great to support your predetermined evaluation and they suck when they they go against it. That stands for each and every one of us.

 

[The Bored Man]

Yawn. More evidence that some hockey bloggers don't know how to interpret statistics.

Mind you, Hall is no defensive gem.

 

[russ99]

LOL. All the Oilers are playing terrible, why pin it on Hall?

It's obvious that Eakins' puck possession/dump system has regressed all our forwards' capacity to create quality scoring chances in the offensive zone.

Maybe we should let these guys take some risks on offense from time to time, just saying. We're already 15-29-5, what's it going to hurt?

 

[CanadianSuperPromise]

Jonathon Willis is an idiot.

Yawn. More evidence that hockey bloggers don't know how to interpret statistics.

Did you understand what the article was saying?

Willis: I do want to make one point which I think most have picked up on. We're comparing Hall's underlying numbers to his fantastic performance last season. I don't think there's any doubt that Hall is an exceptional player, and given that his talent is bound to win the day eventually.

His underlying numbers have dipped sharply this season, and I only see two plausible possibilities: coaching or injury. If it's coaching it might be a case of short-term pain for long-term gain or it might be that Kreuger had a better idea of how to deal with the player than Eakins does. If it's injury, he'll heal. In either case, I doubt we're looking at a permanent drop-off.

 

[russ99]

There is something with Hall? No. There is something with Eakins.

Hall, Eakins 43.2%, Krueger 50.4%

Truth be told Corsi is generally better under Eakins but he sure has some outliers in other stat departments.

The Corsi improvement is due to that we're taking a ton of shots, but they're primarily crappy shots that goalies have no problems with.

 

[Matador]

The thread title and article are completely different. This thread kinda misses the point Willis is making.

 

[Mr Forever]

Did you understand what the article was saying?

Willis is still an idiot

 

[CanadianSuperPromise]

Willis is still an idiot

I don't follow him, but why? Just curious because I don't know.

 

[Gone]

Jonathon Willis is an idiot.

Hardly.

If anything he does show a bit of favoritism to management, yet this piece clearly does the opposite.

 

[Ol' Jase]

I think you interpreted the article wrong.

And there it is!

Lol.

Please share with us all how "the Oilers are terrible when Hall is on the ice" and pointing to SF/60 and SA/60 is to be interpreted?

Tell me, could the dip in SF/60 when Hall is on the ice maybe be a result of Hall and his linemates taking better quality shots? Therefore, resulting in less shots on net and higher offensive output?

However, for that to be a possibility, the offensive numbers would then have to have improved as well, and Hall EV offensive numbers would have to be off the charts.

Oh wait...those numbers are in the article. While I'll be damned!

 

[StoveTopStauffer]

The Corsi improvement is due to that we're taking a ton of shots, but they're primarily crappy shots that goalies have no problems with.

I'd also say its a bit higher as well because we're not always playing heavy west teams like last year and we can actually keep up with shots on some opponents.

 

[Jimmi McJenkins]

Corsi claims that Tom Gilbert is a top-pairing defenseman and is a better player than Andrei Markov.

Until this can be "fixed" to represent reality, there's a flaw there.

 

[Moose Coleman]

And there it is!

Lol.

Please share with us all how "the Oilers are terrible when Hall is on the ice" and pointing to SF/60 and SA/60 is to be interpreted?

Tell me, could the dip in SF/60 when Hall is on the ice maybe be a result of Hall and his linemates taking better quality shots? Therefore, resulting in less shots on net and higher offensive output?

However, for that to be a possibility, the offensive numbers would then have to have improved as well, and Hall EV offensive numbers would have to be off the charts.

Oh wait...those numbers are in the article. While I'll be damned!

Pretty sure exact same stuff about shot quality was trotted out when stats folks were talking about how Eberle's shooting percentage and production was destined to fall. How'd that one turn out again?

 

[Ol' Jase]

The thread title and article are completely different. This thread kinda misses the point Willis is making.

The thread title is a statement right out of the article.

Willis make no attempt to even question whether the dip in shots for is a result of anything other than a regression on Hall's part full stop.

No link to increased shot quality, no mention that the shots could be down a bit because the puck is actually in the Oiler end far more often than not (which is complete folly to attribute to one individual, even though it seems to happen now on a daily basis), no mention of linemates, D partners, etc.

So many missing variables, yet those who question the logic are missing the point. Trust me, there are people that have been asking these same questions from the same people for ages, and the common answer is simply "You just don't get it".

 

[Moose Coleman]

Corsi claims that Tom Gilbert is a top-pairing defenseman and is a better player than Andrei Markov.

Until this can be "fixed" to represent reality, there's a flaw there.

That's not what Corsi claims. Corsi is just a number.

 

[Spawn]

Really the article is showing that the Oilers are a terrible team regardless of who is on the ice, whereas last year Hall was someone who was able be above it.

The only player on this team with a positive corsi is Martin Marincin, and he's only played 11 games. Every single other player on this team on 5v5 gives up more shots than they take.

I'm not a big fan of corsi, but it is interesting that Bob Stauffer has been saying that the teams 5v5 play has improved and uses shot differential to suggest that. Last year the team was 29th in corsi %. This year they are 28th lol.

 

[dustrock]

Well, the fact that the Oilers had a 16 point shot differential with and without Hall last year was pretty crazy - talk about driving possession.

If Willis had just left it as: look, review of shot differential shows that, over a season, it's far better to be outshooting your opponent. Our stats show that the Oilers sucked at this last year, but Hall was the big difference maker, but now this season he isn't a difference maker at all. So the Oilers haven't really improved and Hall has gotten worse (by this statistical measure).

This is concerning because it's hard to understand why, as he has similar linemates and similar QoC, and in fact, Eastern Conference opponents.

Does anyone really have a problem with that? Just wish he had left it at that.

 

[Halibut]

Hardly.

If anything he does show a bit of favoritism to management, yet this piece clearly does the opposite.

I like Willis for the most part even if the stats dont always say what he claims they are saying as with the title in this case. He does really seem to like those old Oilers though. He was always on about how bad Tambellini was as a GM and I kept hammering on his articles at Oilers Nation that Lowe deserved as much blame if not more for hiring the ditherer. Still seems to support Lowe and MacT.

 

[guymez]

Did you understand what the article was saying?

Its pretty clear. Advanced stats are useless and anyone that incorporates them into a discussion or uses them to drive a discussion is an idiot.


Does that help clarify things?

Come on man...there is no room for discussion about this.

Get with the HF program!